def test_load_trained_model_doesnt_train_model_when_already_persisted(self, _, dataset):
dataset.return_value = self.dataset
adapter = Adapter(self.temp_dir)
subject = Core(settings, adapter)
Path(os.path.join(subject.data_path, 'mockedclassifier.pkl')).touch()
model = subject.load_trained_model(self.classifier)
model.fit.assert_not_called()
def test_load_trained_model_trains_model_when_not_persisted(
self, _, dataset):
dataset.return_value = self.dataset
adapter = Adapter(self.temp_dir)
subject = Core(settings, adapter)
subject.load_trained_model(self.classifier)
self.classifier.return_value.fit.assert_called_once_with(self.dataset)
def test_load_trained_model_doesnt_train_model_when_already_persisted(
self, _, dataset):
dataset.return_value = self.dataset
adapter = Adapter(self.temp_dir)
subject = Core(settings, adapter)
Path(os.path.join(subject.data_path, 'mockedclassifier.pkl')).touch()
model = subject.load_trained_model(self.classifier)
model.fit.assert_not_called()
def test_load_trained_model_for_subquota(self):
ClassifierClass, classifier_instance = MagicMock(), MagicMock()
ClassifierClass.return_value = classifier_instance
ClassifierClass.__name__ = 'MonthlySubquotaLimitClassifier'
settings = MagicMock()
settings.UNIQUE_IDS = ['number']
core = Core(settings, self.adapter)
core.load_trained_model(ClassifierClass)
classifier_instance.fit.assert_called_once_with(core.dataset)
def test_load_trained_model_for_subquota(self):
ClassifierClass, classifier_instance = MagicMock(), MagicMock()
ClassifierClass.return_value = classifier_instance
ClassifierClass.__name__ = 'MonthlySubquotaLimitClassifier'
settings = MagicMock()
settings.UNIQUE_IDS = ['number']
core = Core(settings, self.adapter)
core.load_trained_model(ClassifierClass)
classifier_instance.fit.assert_called_once_with(core.dataset)
def test_predict(self):
model = MagicMock()
model.predict.return_value = np.array((1, -1), dtype=np.int)
settings = MagicMock()
settings.UNIQUE_IDS = ['number']
core = Core(settings, self.adapter)
core.predict(model, 'hypothesis')
model.transform.assert_called_once_with(core.dataset)
model.predict.assert_called_once_with(core.dataset)
self.assertFalse(core.suspicions.iloc[0]['hypothesis'])
self.assertTrue(core.suspicions.iloc[1]['hypothesis'])
def test_predict(self):
model = MagicMock()
model.predict.return_value = np.array((1, -1), dtype=np.int)
settings = MagicMock()
settings.UNIQUE_IDS = ['number']
core = Core(settings, self.adapter)
core.predict(model, 'hypothesis')
model.transform.assert_called_once_with(core.dataset)
model.predict.assert_called_once_with(core.dataset)
self.assertFalse(core.suspicions.iloc[0]['hypothesis'])
self.assertTrue(core.suspicions.iloc[1]['hypothesis'])
def test_load_trained_model_for_meal_generalization(self):
ClassifierClass, classifier_instance = MagicMock(), MagicMock()
ClassifierClass.return_value = classifier_instance
ClassifierClass.__name__ = 'MealGeneralizationClassifier'
settings = MagicMock()
settings.UNIQUE_IDS = ['number']
settings.SUPERVISED_MODEL = SUPERVISED_MODEL
core = Core(settings, self.adapter)
core.load_trained_model(ClassifierClass)
classifier_instance.fit.assert_called_once_with(
settings.SUPERVISED_MODEL['MealGeneralizationClassifier'])
def test_load_trained_model_with_pickle(self, joblib, isfile):
isfile.return_value = True
ClassifierClass, classifier_instance = MagicMock(), MagicMock()
ClassifierClass.return_value = classifier_instance
ClassifierClass.__name__ = 'ClassifierMock'
settings = MagicMock()
settings.UNIQUE_IDS = ['number']
core = Core(settings, self.adapter)
core.load_trained_model(ClassifierClass)
expected_path = os.path.join('tmp', 'test', 'classifiermock.pkl')
self.assertFalse(classifier_instance.fit.called)
joblib.load.assert_called_once_with(expected_path)
def test_load_trained_model_with_pickle(self, joblib, isfile):
isfile.return_value = True
ClassifierClass, classifier_instance = MagicMock(), MagicMock()
ClassifierClass.return_value = classifier_instance
ClassifierClass.__name__ = 'ClassifierMock'
settings = MagicMock()
settings.UNIQUE_IDS = ['number']
core = Core(settings, self.adapter)
core.load_trained_model(ClassifierClass)
expected_path = os.path.join('tmp', 'test', 'classifiermock.pkl')
self.assertFalse(classifier_instance.fit.called)
joblib.load.assert_called_once_with(expected_path)
def test_call(self, mocked_predict, mocked_load):
mocked_load.return_value = 'model'
settings = MagicMock()
settings.UNIQUE_IDS = ['number']
settings.CLASSIFIERS = {'answer': 42, 'another': 13}
core = Core(settings, self.adapter)
core.suspicions = MagicMock()
core()
# assert load and predict was called for each classifier
mocked_load.assert_has_calls((call(42), call(13)), any_order=True)
mocked_predict.assert_has_calls(
(call('model', 'answer'), call('model', 'another')),
any_order=True)
# assert suspicions.xz was created
expected_path = os.path.join('tmp', 'test', 'suspicions.xz')
core.suspicions.to_csv.assert_called_once_with(expected_path,
compression='xz',
encoding='utf-8',
index=False)
def test_call(self, mocked_predict, mocked_load):
mocked_load.return_value = 'model'
settings = MagicMock()
settings.UNIQUE_IDS = ['number']
settings.CLASSIFIERS = {'answer': 42, 'another': 13}
core = Core(settings, self.adapter)
core.suspicions = MagicMock()
core()
# assert load and predict was called for each classifier
mocked_load.assert_has_calls((call(42), call(13)), any_order=True)
mocked_predict.assert_has_calls((
call('model', 'answer'),
call('model', 'another')
), any_order=True)
# assert suspicions.xz was created
expected_path = os.path.join('tmp', 'test', 'suspicions.xz')
core.suspicions.to_csv.assert_called_once_with(
expected_path,
compression='xz',
encoding='utf-8',
index=False
)
def test_load_trained_model_trains_model_when_not_persisted(self, _, dataset):
dataset.return_value = self.dataset
adapter = Adapter(self.temp_dir)
subject = Core(settings, adapter)
subject.load_trained_model(self.classifier)
self.classifier.return_value.fit.assert_called_once_with(self.dataset)
def main(target_directory='/tmp/serenata-data'):
adapter = Adapter(target_directory)
core = Core(settings, adapter)
core()
def test_init_without_unique_ids(self):
settings = MagicMock()
settings.UNIQUE_IDS = None
core = Core(settings, self.adapter)
self.assertTrue(core.suspicions.equals(DATAFRAME))
def test_init_with_unique_ids(self):
settings = MagicMock()
settings.UNIQUE_IDS = ['number']
core = Core(settings, self.adapter)
self.assertTrue(core.suspicions.equals(DATAFRAME[['number']]))